1. Field of the Invention
This invention relates to an analog input method which enables minute analog signals obtained from a strain gage, etc., to be read as digital values.
2. Description of the Prior Art
FIG. 1 is a block diagram showing a conventional analog input apparatus shown in Japanese Utility Model Laid-Open No. Sho 58-59218 as an example, wherein numeral 1 is a microprocessor which contains a memory, etc., and reads digital values of an analog-to-digital converter (described below) for data processing; numeral 2 is an analog-to-digital converter which converts a voltage value selected by means of a multiplexer into digital form; numeral 3 is an analog multiplexer for selecting one out of several analog input points; numeral 4A is an oscillator with a duty ratio of 1:1; numeral 5 is a constant-voltage regulated power supply as a bipolar voltage power supply for generating bipolar voltage; numerals 6 and 7 are switches such as mercury contact relays; numerals 11 and 31 are strain gages as bridge measuring devices for measuring distortion caused by external force; numerals 12 and 32 are high input impedance differential amplifiers (hereinafter simply referred to as amplifiers); numerals 13 and 33 are forward amplifiers; numerals 14 and 16 and 15 and 17 are resistors and capacitors respectively making up a low-pass filter; and numerals 34 and 36 and 35 and 37 are resistors and capacitors respectively making up a low-pass filter.
Referring now to the timing chart shown in FIG. 2, operation is described. When the low-to-high transition of an output of the oscillator 4A is made, a contacts of the switches 6 and 7 are turned on, applying +E.sub.0 (V) to the strain gage 11. At the time, assuming that voltages which are minute analog values appearing due to strain gage distortion are v.sub.2 and v.sub.1, distortion containing an offset is .epsilon., and output voltages of the amplifier 12 and forward amplifier 13 are v.sub.+ and V.sub.+, the difference between voltages v.sub.2 and v.sub.1 is as follows: EQU v.sub.2 -v.sub.1 =E.sub.0.(K.epsilon./4) (1)
wherein .epsilon. is distortion obtained from the strain gage and K is a constant determined by the strain gage.
Therefore, the output of the amplifier 12 is as follows: EQU v.sub.+ =g.sub.1 (v.sub.2 -v.sub.1).div..alpha.(v.sub.2 .div.v.sub.1).div..beta..sub.1 (2)
wherein g.sub.1 is the gain of the differential amplifier, in which case the gain is set to about several hundred times higher because the voltages v.sub.1 and v.sub.2 are minute voltages in mV units, .alpha. is the amplification factor for common mode potential which is ideally zero, but actually has a slight value, and .beta..sub.1 is the offset of the differential amplifier.
With the voltage v.sub.+ as input, the following output voltage V.sub.+ is obtained at the forward amplifier 13: EQU V.sub.+ =g.sub.2 v.sub.+ .div..beta..sub.2 =g.sub.2 g.sub.1 (v.sub.2 -v.sub.1).div.g.sub.2 .alpha.(v.sub.2 .div.v.sub.1).div.g.sub.2 .beta..sub.1 .div..beta..sub.2 (3)
wherein g.sub.2 and .beta..sub.2 are the gain and offset of the forward amplifier 13 respectively.
Next, when the high-to-low transition of the output of the oscillator 4A is made, b contacts of the switches and 7 are turned on, applying -E.sub.0 (V) to the strain gage 11. At the time, assuming that voltages appearing due to strain gage distortion are v.sub.2 and v.sub.1 and output voltage of the forward amplifier 13 is V.sub.-, the difference between voltages v.sub.2 and v.sub.1 is as follows: EQU v.sub.1 -v.sub.2 =E.sub.0.(K.epsilon./4) (4)
Therefore, the output voltage of the forward amplifier 13 is as follows: EQU V.sub.- =g.sub.2 g.sub.1 (v.sub.1 -v.sub.2).div.g.sub.2 .alpha.(v.sub.2 .div.v.sub.1).div.g.sub.2 .beta..sub.1 .div..beta..sub.2 (5)
Consequently, from the expressions (1), (4), (3), and (5), the difference between voltages V.sub.+ and V.sub.- is as follows: EQU V.sub.+ -V.sub.- =2g.sub.2 g.sub.1 (v.sub.2 -v.sub.1).div.g.sub.2 .alpha.[(v.sub.2 .div.v.sub.1)-(v.sub.2 .div.v.sub.1)] (6)
Since the second term of expression (6) is near 0, the following is established: EQU .vertline.g.sub.2 .alpha.[(v.sub.2 .div.v.sub.1)-(v.sub.2 .div.v.sub.1)].vertline.&lt;2g.sub.2 g.sub.1 (v.sub.2 -v.sub.1)(7)
Therefore, the offset constituent is removed and distortion .epsilon. can be measured from expressions (1) and (6). An error of the strain gage 11 itself is corrected by the microprocessor from measurement data of the gage. The voltage E.sub.0 of the constant-voltage regulated power supply 5 and the gains g.sub.1 and g.sub.2 are hard to change as compared with the offset. The circuit related to the strain gage 11 is described above; the same also applies to the circuit related to the strain gage 31.
FIG. 3 is a block diagram illustrating a former analog input method, wherein numerals 11-31 are strain gages where the resistance value of one of the arms (for example, R.sub.1) changes with the magnitude of distortion caused by external force, and make up a bridge circuit; numeral 5 is a constant-voltage regulated power supply for applying predetermined supply voltages V.sub.1 and V.sub.2 to the strain gates 11-31; numerals 12-32 are high input impedance differential amplifiers which differentially amplify voltages v.sub.1 and v.sub.2 obtained from two middle points of each of the strain gages 11-31; numeral 3 is a multiplexer for selectively reading one among voltage values v differentially amplified and output by the differential amplifiers 12-32; and numeral 2 is an analog-to-digital converter which converts the analog voltage value v selected at the multiplexer 3 into digital form.
Next, operation is described. When predetermined supply voltages V.sub.1 and V.sub.2 are applied to the strain gages 11-31 from the constant-voltage regulated power supply, the voltages v.sub.1 and v.sub.2 corresponding to the ratio of resistance values of arms R.sub.1 -R.sub.4 are obtained from the two middle points of each of the strain gages 11-31. The relationship between the middle-point voltages v.sub.1 and v.sub.2 and the supply voltages V.sub.1 and V.sub.2 is given from the above-mentioned expression (1) by replacing E.sub.0 of the expression with (V.sub.2 -V.sub.1).
The output voltage values v of the differential amplifiers 12-32, thus amplified, are fed into the multiplexer 3 which then selects one among the voltage values and sends it to the analog-to-digital converter. The selected analog voltage value v is converted into digital form by the analog-to-digital converter 2.
Former analog input apparatus are configured as described above. Thus, when the operating temperature range exceeds 100.degree. C. like 0.degree.-125.degree., when the offset of the amplifier 12 at the first stage changes 1 millivolt or more in terms of gain 1, distortion is small, and it is necessary to raise the gain 1000 times higher, if, for example, the voltage generated due to distortion is made a maximum of 1 millivolt, the gain of the amplifier 12 at the first stage is multiplied by 70, and the gain of the forward amplifier 13 at the second stage is multiplied by 70, the value of g.sub.2 .beta..sub.1 in expression (3) becomes no less than "1 millivolt of offset voltage in terms of gain 1.times.70.times.70=4.9 volts."
Thus, normally the signal also becomes 1 millivolt.times. 70.times.70 and reaches no less than 9.8 volts in total. If the input range of the analog-to-digital converter 2 is .+-.5 volts, that value exceeds this input range. This means that the offset is too large and the gains of the amplifier 12 and forward amplifier 13 cannot be raised. That is, in this case, problems arise such that the gains must be halved, affecting the measurement precision.
When the switches 6 and 7 switch, because of affection of the above-mentioned low-pass filter, rise and fall of output of the forward amplifier 13 become slow and the on/off time of the switches 6 and 7 is slow, thus the read time of analog input becomes slow and the strain gage state also changes, thus problems arise such that when a large number of analog inputs are attempted to be read, it is inevitable to prolong the time because of time constant limits although it should be shorter.
Furthermore, if voltage supplied to the constant-voltage regulated power supply 5 is switching power supplies, output of the power supply 5 also carries spike noise and the above-mentioned resistors 14 and 16 and capacitors 15 and 17 making up the low-pass filter become necessary to remove the spike noise. If the low-pass filter is taken out, precision is adversely affected.
As is obvious from expression (6), it becomes difficult to remove the common mode constituent, and if the common mode gain of the amplifier 12 cannot be made small, precision is adversely affected.
Furthermore, even if the amplifier is adjusted, at room temperature, to set offset .beta. to substantially "0", the amplifier gain g is large, thus even a slight change in the offset voltage causes large affection to occur. Particularly, if the operating temperature range is wide, the fluctuation value with the temperature change is difficult to estimate and measurement precision is adversely affected. If temperatures affect the supply voltages V.sub.1 and V.sub.2 output from the constant-voltage regulated power supply 5, as is obvious from expression (1), it is inevitable to affect measurement precision. The supply voltages applied to the strain gages, V.sub.1 and V.sub.2, are made the same positive and negative voltages and the resistance values are set to substantially the same to make the common mode voltage (v.sub.2 +v.sub.1) substantially "0", thereby ignoring the term .alpha.(v.sub.2 +v.sub.1) of expression (2). However, this measure requires that such setting must be made that the common mode voltage (v.sub.2 +v.sub.1) becomes "0".